(Closes #2636) alter LFRic PSy-layer to lookup nlevels for each kernel

[arporter]

As the title says. I've also renamed 'DynCellIterators' to 'LFRicCellIterators'. It could/should be moved out to domain/lfric too but that can be done as a final step after the bulk of the reviewing is done.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.86%. Comparing base (5833320) to head (084e61f).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #2678   +/-   ##
=======================================
  Coverage   99.86%   99.86%           
=======================================
  Files         353      354    +1     
  Lines       49061    49082   +21     
=======================================
+ Hits        48997    49018   +21     
  Misses         64       64           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[arporter]

This is now ready for a first review. It's a small change but has affected a lot of tests because the name of nlayers in the generated PSy-layer code has changed. It's core LFRic functionality so could be one for @hiker or @TeranIvy.
Integration tests are green.

[hiker]

A lot of work in fixing up the tests, highly appreciated.
I did find the usage of _nlayers_names a bit confusing (in one case a dict, in others a set). Adding a comment or two to indicate why you can be certain that you have the right data structure ("since ... is called in case of ..., we know that _nlayers_name is a dict") would be useful.

It needs to be brought up to current master (conflicts with master). You also indicated that you want to want to move LFRicCellIterators into domain/lfric - I am happy for you to do this now as well, since there is only a comment that needs to be added.

[arporter]

Thanks @hiker. This is ready for another look now. Since I've slightly changed the implementation I haven't performed the split on this iteration. I'll do that once you're happy (and re-launch the integration tests too).

[hiker]

@arporter, the coverage reports one line missing. Not sure if I missed that previously, or if this is a new change??

Otherwise it's looking good, I am happy to do the split and fix the missing test.

[arporter]

I've now done the move and fixed the resulting circular imports (plus the uncovered line).
I also had to write a few tests to get coverage of the 'new' file.
I've triggered the integration tests. Should be ready for another look.

[arporter]

The NEMO5 integration test failed (run crashed) but I don't think that's the fault of this PR.

[hiker]

All changes look good. Integration test failures unrelated to this PR, so I'll proceed to merge.